This adds DT support for the TI DA8xx/OMAP-L1x/AM17xx/AM18xx McBSP driver.
Signed-off-by: NPetr Kulhavy <petr@barix.com>
Reviewed-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Instead of depending on individual SoCs make the edma-pcm depend on the eDMA
dmaengine driver (TI_EDMA).
Update the help text and add DRA7xx family since they have eDMA integrated
as well along with sDMA.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

In case when the dai format is set via the dai_link the format
configuration happens once when the links are probed. If the McASP lose
context after this, the information will be lost and McASP will not going
to work correctly.
To overcome this issue, we save the fmt and set it within hw_params as
well.
Reported-by: NMisael Lopez Cruz <misael.lopez@ti.com>
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NJyri Sarha <jsarha@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The in-driver clock reparenting had been added when we did not had other
means to cleanly set the parent for the fck. Now we can use
assigned-clocks/assigned-clock-parents in DT binding. Print warning when
the fck_parent is present for McASP and recommend the switch to the proper
way to handle the clock selection.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The condition for checking for XDAT being cleared was not correct.

Fixes: 36bcecd0 ("ASoC: davinci-mcasp: Correct TX start sequence")
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org

When McASP is used as TX/RX synchronous (TX side generating clocks for RX
side also) and only capture is used we need to configure the number of TX
slots in order McASP to be able to generate the Frame sync.

Fixes: 9273de1940d9e ("ASoC: davinci-mcasp: Add set_tdm_slots() support")
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Fixes the associations between the tx_mask and rx_mask and the associated
playback / capture streams during setting of the TDM slot. With this
patch in place it is now possible for example to only populate tx_mask
(leaving rx_mask as 0) for output-only codecs to control the TDM slot(s)
the McASP serial port uses for transmit. Before that, this scenario
would incorrectly rely on the rx_mask for this.
Signed-off-by: NAndreas Dannenberg <dannenberg@ti.com>
Reviewed-by: NJyri Sarha <jsarha@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The '\n' at the end of the format string is not needed. It adds an extra
line break when doing
cat /proc/interrupts
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The slot_width is for essentially same thing. Instead of storing
bclk_lrclk_ratio, just store the slot_width. Comments has been updated
accordingly and some variable names changed to more descriptive.
Signed-off-by: NJyri Sarha <jsarha@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Implements set_tdm_slot() callback for mcasp. Channel constraints are
updated according to the configured tdm mask and slots each time
set_tdm_slot() is called. The special case when slot width is set to
zero is allowed and it means that slot width is the same as the sample
width.
Signed-off-by: NJyri Sarha <jsarha@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The TX and RX direction share the same bit clock and frame sync, so
the samplerate must be the same to both directions.
Signed-off-by: NJyri Sarha <jsarha@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The FIFO threshold for McASP should be <=[tx/rx]numevt so the initial value
for the refining should meet this requirement as well.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Use devm_ioremap_resource() instead of open code.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Use resource managed function devm_snd_soc_register_component for
component registration instead of snd_soc_register_component.

Also, remove davinci_vcif_remove as it is now redundant.
Signed-off-by: NVaishali Thakkar <vthakkar1994@gmail.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Find the configured DMA controller by asking for a DMA channel in the
probe phase and releasing it right after. The controller device can be
found via the dma_chan struct and the controller is recognized from
the compatible property of its device node. The patch assumes EDMA if
there is no device node.
Signed-off-by: NJyri Sarha <jsarha@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The default state when serializers are in inactive slots is Hi-Z.
In some cases, there are no additional components driving the data
lines to a safe state so they might have noise.

While in inactive slots, the McASP AXR pins configured as outputs
can be driven low through the serializer pin drive mode setting
(DISMOD) to prevent such noise.
Signed-off-by: NMisael Lopez Cruz <misael.lopez@ti.com>
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

pm_runtime_enabled() will only tell if the pm runtime has been enabled for
the device, which is done at probe time but will not tell the actual power
state of the device.
pm_runtime_active() provides this information.
This patch fixes a kernel crash when doing suspend when McASP is not
active.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

There is no need to copy the list of all supported sample-rates.
Finding the supported endpoints within the current range is enough
(see snd_interval_list()).
Signed-off-by: NJyri Sarha <jsarha@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Set channel count constraints for multiple serializers case. On McASP
the active channels mask is the same for all the serializers. With the
current implementation this means that if more than one serializers is
used, all TDM slots have to be active on all serializers. The patch
sets the channel count constraints according to number of RX and TX
serializers.
Reported-by: NMisael Lopez Cruz <misael.lopez@ti.com>
Signed-off-by: NJyri Sarha <jsarha@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The McASP driver currently always sends as many slots or channels to a
i2s-wire as there are configured tdm_slots (see mcasp_i2s_hw_param()).
Thus the BLCK rate does not depend on the amount of channels, just the
configure amount of tdm-slots.
Reported-by: NMisael Lopez Cruz <misael.lopez@ti.com>
Signed-off-by: NJyri Sarha <jsarha@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The dapm field of the snd_soc_codec struct will eventually be removed
(replaced with the DAPM context from the component embedded inside the
CODEC). Replace its usage with the card's DAPM context. The idea is that
DAPM is hierarchical and with the card at the root it is possible to access
widgets from other contexts through the card context.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

Passing &mcasp->ruledata[dir] to snd_pcm_hw_rule_add() is not correct since
commit:
7b3d165a ASoC: davinci-mcasp: Index ruledata in drvdata with substream->stream
now sets up the struct based on the substream->stream (0 or 1) while we pass
a pointer which we take with dir (1 or 2). This will lead kernel crash.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The serializer direction definitions runs from 1 to 2, which does not
suite the purpose. The substream->stream is perfect for the purpose
and should have been used from the beginning.
Signed-off-by: NJyri Sarha <jsarha@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

As davinci card gets registered using 'devm_' api
there is no need to unregister the card in 'remove'
function.
Hence drop the 'remove' function.

Fixes: ee2f615d (ASoC: davinci-evm: Add device tree binding)
Signed-off-by: NManish Badarkhe <manishvb@ti.com>
Signed-off-by: NJyri Sarha <jsarha@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org

Set rule constraints to allow only combinations of sample-rate,
sample-format, and channels counts that can be played/captured with
reasonable sample-rate accuracy.

The logic with tdm-slots and serializers (=i2s data wires) goes like
this: The first wire will take all channels up to number of tdm-slots,
before following wires (if any) are used. If the first wire is used
fully, the remaining wires share the same clocks and the divider can
be calculated for the first wire.

Also, takes the number of tdm-slots into account when implicitly
selecting the BLCK divider.
Signed-off-by: NJyri Sarha <jsarha@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Introduced by commit:
6afda7f5 ASoC: davinci-mcasp: Allow complete shutdown of McASP when not in use

I'm really sorry for this...
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Rearrange the pm_runtime_get/put_sync calls so the IP will be turned off
when it is not in use.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

All DAI drivers has been converted to use edma-pcm instead of davinci-pcm
and the driver can be removed from the tree.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The edma-pcm performs as good as the old davinci-pcm and it's use does not
require the 'ping-pong' mode of davinci-pcm, which was introduced to
overcome under/over flow issues when using davinci-pcm.

Keep the SND_DAVINCI_SOC config option to select the SND_EDMA_SOC to avoid
regression in audio support.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The edma-pcm can replace the old davinci-pcm as platform driver.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The edma-pcm can replace the old davinci-pcm as platform driver.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

edma-pcm going to replace davinci-pcm as platform driver for daVinci
platform.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Support for setups where codec is bitclock slave and frame master.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

One of the McASP instances in DM646x line of DMSoC only supports DIT mode.
This means that the given IP does not have support for rx and all the rx
related resources are missing, like irq and DMA request.
The driver should not fail if any or all  of the RX resource is missing
when the op_mode is set to DIT mode.
Since RX is not possible in DIT mode, we can just ignore the rx resources
when the McASP is used in DIT mode.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

On DA830 devices McASP0,1 and 2 shares a single combined interrupt request
line.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

There is no need for them to be included.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The machine driver has no business with the underlying dma.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

An earlier bug fix of mine made the SND_DM365_VOICE_CODEC symbol
tristate to avoid creating an undefined reference from the
davinci-vcif.c driver to the davinci_soc_platform_register
function that may be in a module.

However, this may now lead to a different error on randconfig
kernels:

"warning: SND_DM365_VOICE_CODEC creates inconsistent choice state"

This happens because we now have a choice statement with
one bool and one tristate option, and the latter might not
support being set to 'y' because of dependencies.

This new change turns the other option into 'tristate' as well,
which avoids the problem.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Fixes: 19926c6d ("ASoC: davinci: vcif must be a module if SND_DAVINCI_SOC is")
Signed-off-by: NMark Brown <broonie@kernel.org>

Some SoC, like da850/OMAP-L138 uses one common interrupt request for TX/RX
events.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

An underrun (playback) event occurs when the serializer transfer
data from the XRBUF buffer to the XRSR shift register, but the
XRBUF hasn't been filled. Similarly, the overrun (capture) event
occurs when data from the XRSR shift register is transferred to
the XRBUF but it hasn't been read yet.

These events are handled as XRUN events that cause the pcm to stop.
The stream has to be explicitly restarted by the userspace which
ensures that after stopping/starting McASP the data transfer is
aligned with DMA. The other possibility was to internally stop and
start McASP without DMA even knowing about it.
Signed-off-by: NMisael Lopez Cruz <misael.lopez@ti.com>
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>